EP0019358B1 - Hierarchisches Datenspeichersystem - Google Patents

Hierarchisches Datenspeichersystem Download PDF

Info

Publication number
EP0019358B1
EP0019358B1 EP80301192A EP80301192A EP0019358B1 EP 0019358 B1 EP0019358 B1 EP 0019358B1 EP 80301192 A EP80301192 A EP 80301192A EP 80301192 A EP80301192 A EP 80301192A EP 0019358 B1 EP0019358 B1 EP 0019358B1
Authority
EP
European Patent Office
Prior art keywords
store
data
location
address
item
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired
Application number
EP80301192A
Other languages
English (en)
French (fr)
Other versions
EP0019358A1 (de
Inventor
Philip Broughton
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fujitsu Services Ltd
Original Assignee
Fujitsu Services Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Fujitsu Services Ltd filed Critical Fujitsu Services Ltd
Publication of EP0019358A1 publication Critical patent/EP0019358A1/de
Application granted granted Critical
Publication of EP0019358B1 publication Critical patent/EP0019358B1/de
Expired legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
    • G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
    • G06F12/0804—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with main memory updating
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
    • G06F12/10—Address translation
    • G06F12/1027—Address translation using associative or pseudo-associative address translation means, e.g. translation look-aside buffer [TLB]
    • G06F12/1045—Address translation using associative or pseudo-associative address translation means, e.g. translation look-aside buffer [TLB] associated with a data cache

Definitions

  • This invention relates to hierarchical data storage systems.
  • a hierarchical data storage system basically comprises a first, relatively large store, in conjunction with a second store which is smaller but faster than the first store.
  • the second store is arranged to hold copies of data items which have recently been used and therefore have a high probability of being required again in the near future.
  • a high proportion of data retrieval requests can be satisfied simply by reading the data from the second store, without the need for access to the first store.
  • the system can achieve a fast average data access speed, ideally approaching that of the second store, but with a low average cost per bit of storage, approaching that of the first store.
  • the object of the present invention is to alleviate the problems of (a) and (b) above without the necessity of providing an additional store for the storage of all the items of data which have been modified and without any need for priority resolution hardware.
  • a hierarchical data storage system comprising a first store and a second store having a faster acess time but a smaller number of data storage locations than the first store and a further store for entries relating to data items in said second store which have been modified relative to corresponding data items in the first store, the entries of the further store being used for replacing data items in the second store by writing into the first store those items which have not previously been transferred; characterised in that each location of the second store has a changed bit indicating in its set state that the data item in that location has been modified and in its reset state that the data item has the same value as the corresponding item in the first store; in that entries in the further store are organised as a queue of entries; in that when a modified item of data is to be written into one of the storage locations of the second store selected by an address generator the state of the changed bit is tested for an indication that the data item in the addressed storage location has not been modified, the new or modified item of data is written in that storage location unless the changed bit of that location is already set
  • Figure 1 is a schematic block diagram of a data processing system including a cache store
  • the system comprises a data processing unit 1, a main store 2, and a cache store 3 situated between the processing unit and the main store.
  • the main store 2 is a relatively large store holding, for example, 64K data items. (These items may for example be words, or blocks of words). Every data item has a real address specifying its location within the main store.
  • the cache store 3 is substantially smaller than the main store, but has a substantially faster access time.
  • the processing unit 1 Whenever the processing unit 1 requires to read a particular data item, it applies the virtual address of that item to the cache store 3 and issues a READ command. A check is then made to find out whether that item is currently resident in the cache and, if it is, the item is read out without any need for a main store access. However, if the item is not in the cache, the virtual address is translated into the real address of the item and a main store access is initiated so as to read the item from the main store and an attempt is made to place it in the cache. However, if the location into which the item would be placed is already occupied by another item whose changed bit is set, this location cannot be overwritten since it contains the only up-to-date copy of that other item. Hence, the newly retrieved data item by-passes the cache to the processing unit and is not retained by the cache.
  • the processing unit When the processing unit requires to modify a data item, it applies the virtual address of the item to the cache 3 along with the new value of the data, and issues a WRITE command. An attempt is then made to write the new value into the cache. A check is first made to find out whether the location into which the modified item would be placed is already occuplied by the existing value of the same item. If so, the new value is written into the location, overwriting the existing value, and the changed bit is set (if it is not already set). Likewise, if the location is occupied by a different item whose changed bit is not set, the modified item is written into the location, overwriting the existing item, and then the changed bit is set. However, if the location is occupied by a different item whose changed bit is set, that item cannot be overwritten, and so the modified item must be transferred immediately to the main store, without being entered into the cache.
  • the changed list has 32 locations and entries in the changed list are effectively organised as a queue. Whenever a data item is modified in the cache, the cache address of that item is inserted into the changed list at the tail of the queue (unless the changed bit of that item is already set, indicating that the item has already been modified and hence the cache address of that item is already in the queue). When the changed list becomes full, the entry at the head of the queue is read out (thus reducing the number of entries remaining in the queue to 31) and is used to address the cache. The addressed item and its real address are then read out of the cache and a main store access is initiated, so as to transfer the item to the main store. The changed bit of the item is reset.
  • the cache store comprises a random access memory 10 having 1024 individually addressable locations. Each location contains the following:-
  • the input virtual address from the processing unit is applied to an address translation unit 12 which translates it into the corresponding real address, which is applied to the input of the RA section of the cache.
  • the virtual address is also applied to the VA section of the cache, and to the input of a hashing circuit 13.
  • the hashing circuit performs a pseudo-random transformation on the virtual address to produce a 10-bit hash code for addressing the cache.
  • the hashing circuit may, for example, consist of a set of ten exclusive-OR gates, each of which receives two bits of the virtual address.
  • the address input of the cache is connected to a multiplexer 15, which selects either the output of the hashing circuit 13 or the output of the changed list 4.
  • the output of the hashing circuit is also applied to the data input of the changed list 4.
  • the changed list is addressed by a 5-bit cyclic counter REJB.
  • the system also includes an address buffer 17 and a data buffer 18.
  • the input of the address buffer is connected to the output of a multiplexer 19 which selects either the output of the address translation unit 12, or the output of the RA section of the cache 10.
  • the input of the data buffer 18 is connected to the output of a multiplexer 20 whose inputs are connected to an input data path 21, which receives data from the processing unit, and to the output of the DATA section of the cache.
  • the data path 21 is also connected to the input of the DATA section of the cache.
  • the output of the VA section of the cache is applied to a comparator 22 where it is compared with the input virtual address received on path 11.
  • the output of this comparator provides a signal INCACHE which is true if the input virutal address matches the output from the cache. This signal therefore indicates that the data item whose virtual address is supplied on path 11 is currently resident in the cache.
  • the system also includes three bistables referred to as the request flag REQ, the decrement flag DECR, and the overflow flag REJBOV, the purpose of which will be explained below.
  • this shows the logic circuit which controls the execution of a WHITE command.
  • the logic circuit comprises two bistables 25, 26 which respectively define two control states. When a WRITE command is received, bistable 25 is set, placing the logic in the first control state.
  • a signal W1 is produced, which sets the multiplexers 15, 19 and 20 so that each selects its input labelled "0" in Figure 2.
  • the address input of the cache 10 receives the hash code from the hashing circuit 13
  • the address buffer 17 receives the real address from the address translation unit 12, and the data buffer receives the modified data from path 21.
  • the signal W1 also clocks the changed list 4, causing the hash code to be written into the location currently addressed by the counter REJB.
  • the signal W1 also resets the decrement flag DECR.
  • a further delay (D2) now follows, to allow the addressing of the cache 10 to be completed and the contents of the addressed location to be read out.
  • the signal INCACHE is available from comparator 22, to indicate whether the data item in question is currently resident in the cache.
  • the changed bit CH of the addressed item is also available.
  • an OR gate 27 is enabled, which in turn enables an AND gate 29, producing a signal W4.
  • This signal clocks the VA, RA and DATA portions of the cache 10, causing the virtual address VA, real address from the address translation unit and the modified data from path 21 to be written into the currently addressed location of the cache.
  • This signal clocks the buffers 17 and 18, so as to write the modified data from path 21 into the data buffer 18 and the corresponding real address into the address buffer 17.
  • the signal W5 also sets the request flag REQ. This initiates transfer of the contents of the buffer 18 into the location of the main store specified by the contents of the address buffer 17.
  • the bistable 25 is reset, terminating the first control state, and the bistable 26 is set, initiating the second control state.
  • a signal W6 is produced, which sets the multiplexers 15, 19 and 20 so that they now select the inputs labelled "1"in Figure 2.
  • a delay (D4) now follows, allowing the cache to be addressed, via multiplexer 15, by the output of the changed list.
  • an AND gate 33 is enabled, producing a signal W9.
  • This signal indicates that there are currently 32 modified data items in the cache and hence one of these items must be transferred to the main store to reduce the number to 31, which is the maximum number of modified items allowed to remain in the cache awaiting transfer.
  • the signal W9 therefore clocks the buffers 17 and 18, causing the currently addressed data item in the cache to be written into the data buffer 18 and its real address to be written into the address buffer 17.
  • the signal W9 also resets the changed bit of the currently addressed data item, and sets the request flag REQ, initiating the transfer of the data item to the main store.
  • the bistable 26 After a further delay (D5), the bistable 26 is reset, terminating the second control state. In the case where the request flag REQ has been set, it remains set until the data transfer is completed, whereupon it is reset by a signal from the main store.
  • this shows the logic for controlling execution of a CLEAR command.
  • the execution consists of a number of cycles. In each cycle, the currently addressed location of the changed list is read out and used to address the cache, and the addressed item is then transferred to the main store. The counter REJB is decremented and the cycle is then repeated until all modified items have been transferred. The number of cycles to be performed depends on whether the overflow flag REJBOV has been set.
  • the counter REJB has completed at least one full counting cycle. Hence, there must be 31 modified data items in the cache, and the transfer cycle must be performed 31 times. This is controlled by means of another 5-bit counter RSC which is initially set to 1 and is incremented each time a data item is transferred.
  • RSC 3 1 . This condition is detected by an AND gate 41.
  • the outputs of the two AND gates 40, 41 are combined in an OR gate 42 to produce a signal which indicates that all the data items have been transferred.
  • a bistable 43 When a CLEAR command is received, a bistable 43 is set. This produces a signal C1 which causes the multiplexers 15, 19 and 20 to be switched so as to select the inputs labelled ' ' 1' ' . The signal C1 also resets the counter RSC to its initial value "1 ".
  • a signal C2 is produced. This signal clocks the buffers 17 and 18 so as to write the currently addressed data item from the cache into the data buffer 18 and its real address into the address buffer 17.
  • the signal C2 also resets the changed bit in the currently addressed location, and sets the request flag REQ, initiating the transfer of the data item to the main store.
  • the circuit now waits for the transfer of the data item to the main store to be completed, and the request flag REQ to be reset.
  • an AND gate 46 is enabled, and the bistable 45 is reset. Enabling of the AND gate 46 results in the signal C2 being produced again, initiating the transfer of the next data item.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Memory System Of A Hierarchy Structure (AREA)

Claims (8)

1. Hierarchische Datenspeicheranordnung mit einem ersten Speicher (2) und einem zweiten Speicher (10), der eine höhere Zugriffsgeschwindigkeit, jedoch eine geringere Anzahl von Datenspeicherplätzen besitzt als der erste Speicher, und mit einem weiteren Speicher (4) für Einträge, die sich auf Datenausdrücke im zweiten Speicher beziehen, die in bezug auf entsprechende Datenausdrücke im ersten Speicher modifiziert worden sind, wobei die Einträge des weiteren Speichers verwendet werden, um Datenausdrücke im zweiten Speicher dadurch zu ersetzen, daß die Ausdrücke, die vorher nicht übertragen worden sind, in den ersten Speicher eingeschrieben werden, dadurch gekennzeichnet, daß jeder Speicherplatz des zweiten Speichers ein geändertes Bit (CH) besitzt, das im gesetzten Zustand anzeigt, daß der Datenausdruck in diesem Speicherplatz modifiziert worden ist, und das im rückgesetzten Zustand anzeigt, daß der Datenausdruck den gleichen Wert wie der entsprechende Ausdruck im ersten Speicher hat, daß Einträge in den weiteren Speicher (4) als eine Reihe von Einträgen organisiert sind, daß dann, wenn ein modifizierter Datenausdruck in einen der Speicherplätze des zweiten Speichers eingeschrieben werden soll, der durch einen Adressengenerator (13) ausgewählt wird, der Zustand des geänderten Bits (CH) auf eine Anzeige, daß der Datenausdruck in dem adressierten Speicherplatz nicht modifiziert worden ist, geprüft wird (durch Gatter 27, 29), der neue oder modifizierte Datenausdruck in den Speicherplatz eingeschrieben wird, wenn das geänderte Bit dieses Speicherplatzes nicht bereits gesetzt ist, und die Adresse dieses Speicherplatzes durch den Adressengenerator (13) als ein Eintrag am Ende der Reihe in den weiteren Speicher (4) eingeschrieben wird, und daß ein Zähler (REJB) die Anzahl von Einträgen in den weiteren Speicher anzeigt und, wenn der weitere Speicher gefüllt ist, der Zähler am Beginn der Reihe ein Signal erzeugt, um den entsprechenden Speicherplatz des zweiten Speichers so zu adressieren, daß der Datenausdruck in diesem adressierten Speicherplatz auf den ersten Speicher übertragen wird und das geänderte Bit dieses Speicherplatzes rückgesetzt wird.
2. Anordnung nach Anspruch 1, dadurch gekennzeichnet, daß die Anzahl von Einträgen in den weiteren Speicher (4) wesentlich kleiner ist als die Anzahl von Speicherplätzen im zweiten Speicher.
3. Anordnung nach Anspruch 1 oder 2, dadurch gekennzeichnet, daß der weitere Speicher (4) ein Speicher mit direktem Zugriff ist, daß der Zähler (REJB) mit dem Adresseneingang des Speichers verbunden ist, und daß Mittel vorgesehen sind, um den Zählter weiterzuschalten, wenn in den Speicher (4) eingeschrieben wird.
4. Anordnung nach einem der vorausgehenden Ansprüche, gekennzeichnet durch Mittel zum serienweisen, selbsttätigen Auslesen aus der Reihe und zum Verwenden dieser Einträge für das Adressieren des zweiten Speichers (1) um alle modifizierten Datenausdrücke, die gerade im zweiten Speicher vorhanden sind, in den ersten Speicher (2) zu übertragen.
5. Anordnung nach einem der vorausgehenden Ansprüche, gekennzeichnet durch eine Kontrollsummenschaltung (13) zur Durchführung einer pseudowillkürlichen Transformation an einer virtuellen Eingangsadresse (VA), um eine Kontrollsummenadresse zu erzeugen, und einen Multiplexer (15) zur Auswahl entweder der Kontrollsummenadresse oder des Ausganges des weiteren Speichers (4) zum Adressieren des zweiten Speichers (10).
6. Anordnung nach Anspruch 5, gekennzeichnet durch Mittel (29), die dann, wenn Daten in einen Speicherplatz des zweiten Speichers eingeschrieben werden, bewirken, daß die virtuelle Eingangsadresse (VA) in den gleichen Speicherplatz eingeschrieben wird.
7. Anordnung nach Anspruch 6, gekennzeichnet durch Mittel (22) zum Vergleichen der virtuellen Eingangsadresse (VA) mit der virtuellen Adresse, die in einem Speicherplatz des zweiten Speichers (10) gespeichert ist.
8. Anordnung nach einem der vorausgehenden Ansprüche, gekennzeichnet durch eine Adressenumsetzungseinheit (12) zum Umsetzen der virtuellen Adresse (VA) in eine echte Adresse (RA), und Mittel (29), die dann, wenn Daten in einen Speicherplatz des zweiten Speichers eingeschrieben werden, bewirken, daßdie echte Adresse in den gleichen Speicherplatz eingeschrieben wird.
EP80301192A 1979-05-09 1980-04-15 Hierarchisches Datenspeichersystem Expired EP0019358B1 (de)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB7916021 1979-05-09
GB7916021 1979-05-09

Publications (2)

Publication Number Publication Date
EP0019358A1 EP0019358A1 (de) 1980-11-26
EP0019358B1 true EP0019358B1 (de) 1984-07-11

Family

ID=10505030

Family Applications (1)

Application Number Title Priority Date Filing Date
EP80301192A Expired EP0019358B1 (de) 1979-05-09 1980-04-15 Hierarchisches Datenspeichersystem

Country Status (5)

Country Link
US (1) US4322815A (de)
EP (1) EP0019358B1 (de)
AU (1) AU538533B2 (de)
DE (1) DE3068498D1 (de)
ZA (1) ZA802368B (de)

Families Citing this family (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4382278A (en) * 1980-06-05 1983-05-03 Texas Instruments Incorporated Hierarchial memory system with microcommand memory and pointer register mapping virtual CPU registers in workspace cache #4 and main memory cache
US4464713A (en) * 1981-08-17 1984-08-07 International Business Machines Corporation Method and apparatus for converting addresses of a backing store having addressable data storage devices for accessing a cache attached to the backing store
AU542447B2 (en) * 1982-02-27 1985-02-21 Fujitsu Limited System for controlling key storage unit
JPS5994289A (ja) * 1982-11-22 1984-05-30 Hitachi Ltd 記憶制御方式
WO1984002799A1 (en) * 1982-12-30 1984-07-19 Ibm A hierarchical memory system including separate cache memories for storing data and instructions
US4897783A (en) * 1983-03-14 1990-01-30 Nay Daniel L Computer memory system
DE3377435D1 (en) * 1983-07-27 1988-08-25 Ibm Deutschland Digital memory
US4747044A (en) * 1984-08-23 1988-05-24 Ncr Corporation Direct execution of software on microprogrammable hardware
US4991081A (en) * 1984-10-31 1991-02-05 Texas Instruments Incorporated Cache memory addressable by both physical and virtual addresses
US4737909A (en) * 1985-04-01 1988-04-12 National Semiconductor Corp. Cache memory address apparatus
US4875155A (en) * 1985-06-28 1989-10-17 International Business Machines Corporation Peripheral subsystem having read/write cache with record access
US4803651A (en) * 1986-02-10 1989-02-07 Ibm Corporation Method for encoding and decoding acyclic, linked-list data structures for efficient storage, retrieval and searching
US5276826A (en) * 1988-01-04 1994-01-04 Hewlett-Packard Company Apparatus for transforming addresses to provide pseudo-random access to memory modules
NL8800158A (nl) * 1988-01-25 1989-08-16 Philips Nv Computersysteem voorzien van een hierarchisch georganiseerd geheugen.
AU2278688A (en) * 1988-02-16 1989-08-17 Sun Microsystems, Inc. Distributed cache architecture
GB8814077D0 (en) * 1988-06-14 1988-07-20 Int Computers Ltd Data memory system
JP2695017B2 (ja) * 1989-11-08 1997-12-24 富士通株式会社 データ転送方式
US5226133A (en) * 1989-12-01 1993-07-06 Silicon Graphics, Inc. Two-level translation look-aside buffer using partial addresses for enhanced speed
JPH0821003B2 (ja) * 1992-08-07 1996-03-04 インターナショナル・ビジネス・マシーンズ・コーポレイション コンピュータ・キャッシュ・システム用の加算器／ハッシュ回路
US5502833A (en) * 1994-03-30 1996-03-26 International Business Machines Corporation System and method for management of a predictive split cache for supporting FIFO queues
US5555379A (en) * 1994-07-06 1996-09-10 Advanced Micro Devices, Inc. Cache controller index address generator
GB2293670A (en) * 1994-08-31 1996-04-03 Hewlett Packard Co Instruction cache
JPH10320274A (ja) * 1997-03-19 1998-12-04 Toshiba Corp キャッシュフラッシュ装置及び同装置を備えた計算機システム、記録媒体
US20120102261A1 (en) * 2009-07-07 2012-04-26 Harley Burger Systems and Methods for Tiered Non-Volatile Storage

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3976983A (en) * 1974-02-15 1976-08-24 U.S. Philips Corporation Bipolar programmable read only memory with fusible links
US4045784A (en) * 1975-01-10 1977-08-30 Nippon Electric Co., Ltd. Programmable read only memory integrated circuit device

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3292152A (en) * 1962-09-17 1966-12-13 Burroughs Corp Memory
US3588839A (en) * 1969-01-15 1971-06-28 Ibm Hierarchical memory updating system
US3609665A (en) * 1970-03-19 1971-09-28 Burroughs Corp Apparatus for exchanging information between a high-speed memory and a low-speed memory
US3878513A (en) * 1972-02-08 1975-04-15 Burroughs Corp Data processing method and apparatus using occupancy indications to reserve storage space for a stack
US3818461A (en) * 1972-04-10 1974-06-18 Litton Systems Inc Buffer memory system
US3868644A (en) * 1973-06-26 1975-02-25 Ibm Stack mechanism for a data processor
NL7317545A (nl) * 1973-12-21 1975-06-24 Philips Nv Geheugensysteem met hoofd- en buffergeheugen.
US4020466A (en) * 1974-07-05 1977-04-26 Ibm Corporation Memory hierarchy system with journaling and copy back
US4056845A (en) * 1975-04-25 1977-11-01 Data General Corporation Memory access technique
GB1533768A (en) * 1975-12-18 1978-11-29 Ibm Data processing apparatus
US4008460A (en) * 1975-12-24 1977-02-15 International Business Machines Corporation Circuit for implementing a modified LRU replacement algorithm for a cache
US4053948A (en) * 1976-06-21 1977-10-11 Ibm Corporation Look aside array invalidation mechanism
US4157586A (en) * 1977-05-05 1979-06-05 International Business Machines Corporation Technique for performing partial stores in store-thru memory configuration
US4229789A (en) * 1977-12-22 1980-10-21 Ncr Corporation System for transferring data between high speed and low speed memories
US4195340A (en) * 1977-12-22 1980-03-25 Honeywell Information Systems Inc. First in first out activity queue for a cache store

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3976983A (en) * 1974-02-15 1976-08-24 U.S. Philips Corporation Bipolar programmable read only memory with fusible links
US4045784A (en) * 1975-01-10 1977-08-30 Nippon Electric Co., Ltd. Programmable read only memory integrated circuit device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
R.E. Matick, "Computer Storage Systems and Technology", 1977, pages 591-592, 633-638 *

Also Published As

Publication number Publication date
US4322815A (en) 1982-03-30
EP0019358A1 (de) 1980-11-26
AU538533B2 (en) 1984-08-16
AU5822080A (en) 1980-11-13
DE3068498D1 (en) 1984-08-16
ZA802368B (en) 1981-04-29

Similar Documents

Publication Publication Date Title
EP0019358B1 (de) Hierarchisches Datenspeichersystem
KR100278328B1 (ko) 캐시 미스 버퍼
US5454093A (en) Buffer bypass for quick data access
US4298929A (en) Integrated multilevel storage hierarchy for a data processing system with improved channel to memory write capability
EP0009938B1 (de) Computersysteme mit Hochgeschwindigkeits-Cache-Speichern
US4442488A (en) Instruction cache memory system
US5230045A (en) Multiple address space system including address translator for receiving virtual addresses from bus and providing real addresses on the bus
US5123101A (en) Multiple address space mapping technique for shared memory wherein a processor operates a fault handling routine upon a translator miss
US4400774A (en) Cache addressing arrangement in a computer system
JPS624745B2 (de)
US4323968A (en) Multilevel storage system having unitary control of data transfers
US4912631A (en) Burst mode cache with wrap-around fill
US4774687A (en) Advanced store-in system for a hierarchy memory device
US4445191A (en) Data word handling enhancement in a page oriented named-data hierarchical memory system
EP0173909B1 (de) LRU-Markiersteuerungsschaltung für look-aside Puffer
US4424564A (en) Data processing system providing dual storage of reference bits
JPS623354A (ja) キヤツシユメモリ・アクセス方式
JP2700147B2 (ja) 命令キャッシュ・フラッシュ・オン・ｒｅｉ制御
CA1328026C (en) Apparatus and method for enhanced virtual to real address translation for accessing a cache memory unit
JPH05216767A (ja) エントリ置き換え制御方式
JP2586160B2 (ja) アドレス変換機構
JP2864548B2 (ja) 命令キャッシュ装置
JPH04353949A (ja) キャッシュメモリ制御方式
JPH0677240B2 (ja) キャッシュメモリ制御回路
EP0346918A2 (de) Vorrichtung für Cache-Speicherverwaltung

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Designated state(s): DE FR GB IT SE

17P Request for examination filed

Effective date: 19810521

ITF It: translation for a ep patent filed

Owner name: DR. ING. A. RACHELI & C.

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Designated state(s): DE FR GB IT SE

REF Corresponds to:

Ref document number: 3068498

Country of ref document: DE

Date of ref document: 19840816

ET Fr: translation filed
PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed
ITTA It: last paid annual fee
PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 19910619

Year of fee payment: 12

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Effective date: 19930101

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: SE

Payment date: 19930416

Year of fee payment: 14

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 19930429

Year of fee payment: 14

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SE

Effective date: 19940416

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Effective date: 19941229

EUG Se: european patent has lapsed

Ref document number: 80301192.3

Effective date: 19941110

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 19990303

Year of fee payment: 20

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF EXPIRATION OF PROTECTION

Effective date: 20000414

REG Reference to a national code

Ref country code: GB

Ref legal event code: PE20

Effective date: 20000414